An Early Pliocene relative sea level record from Patagonia (Argentina)

We report ciao a geological unit surveyed and dated in central Patagonia, Argentina (Camarones town, San Jorge Gulf). The unit was interpreted as representative of an intertidal environment and dated to the Early Pliocene (4.69-5.23 Ma) with strontium isotope stratigraphy. The elevation of this unit was measured with differential GPS at ca. 36 m above present-day sea level. Considering modern tidal ranges, it was possible to constrain paleo relative sea level within ±2.5m (1σ). We use glacial isostatic adjustment models and estimates of vertical land movement to calculate that, when the Camarones intertidal sequence was deposited, global mean sea level was 28.4 ± 11.7m above present. This estimate matches those derived from analogous Early Pliocene sea level proxies in the Mediterranean Sea and South Africa. Evidence from these three locations indicates that Early Pliocene sea level may have exceeded 20m above its present level. Such high global mean sea level values imply an ice-free Greenland, a significant melting of West Antarctica, and a contribution of East Antarctica to global mean sea level.

. This caused a switch from subsidence to uplift, and the 100 Patagonia region underwent a moderate but continuous uplift. 26 . 101 Along the coastlines of Central Patagonia, several levels of pa-102 leo shorelines above modern sea level were already noted by 103 Charles Darwin in his Beagle voyage 27 , and were the subject 104 of more than 150 years of research (see Supplementary Infor-105 mation for details). Studies in Central Patagonia include coastal 106 sequences of Holocene 28;29 , Pleistocene 30;31;32 and Pliocene-107 to-Miocene 33;34 age. Among the latter, Del Río et al 34 dated 108 Early Pliocene mollusks from marine deposits few hundreds of 109 kilometers south of the study area described in this study (see 110 Supplementary Information for details). Italian geologist Feruglio 36 identified an elevated marine ter-117 race along a roadcut carved on the main road leading into the 118 town of Camarones that he tentatively attributed to the Pliocene. 119 A recent study 30 confirmed the elevation of this terrace at ca.
glomerates with rounded pebbles and shells. In particular, the 145 uppermost part of this unit contains a dense faunal assemblage 146 in the form of a shellbed, where we recognized 15 different 147 species of bivalves and 11 species of gastropods (see  mentary Information for details). The bivalve shells are mostly 149 intact and sometimes with paired valves (articulated), but not 150 in living position. This unit was interpreted as representative 151 of a foreshore environment, i.e. the intertidal zone. The same 152 unit has been identified at the Caprock section, at roughly the 153 same elevation. The elevation of Unit Cp was measured at two 154 points at both Roadcut and Caprock (Table 1). From these mea-155 surements, we calculate that Unit Cp has an average elevation 156 of 36.2 ± 0.5m (1σ) above the GEOIDEAR16 geoid 39 , which 157 approximates present sea level. Using modern tidal values 35 , 158 and assuming no post-depositional movement, we calculate that 159 the two outcrops in the area of Camarones are indicative of a 160 paleo RSL at 36.2 ± 2.5m (1σ) above present (see Methods for 161 details).

162
Age. Three oyster shells from Roadcut and Caprock were 163 analyzed by Strontium Isotope Stratigraphy (SIS) relative dating 164 techniques. Using sequential leaching to target the least altered 165 inner carbonate of each shell (Sandstrom et al., under review), 166 we obtained multiple SIS ages on three different shells (one 167 from Caprock and two from Roadcut). The shells yielded an 168 age range of 4.69-5.23Ma (n=6, 2σ S EM ) (see Methods and 169 Supplementary Information for details). Glacial Isostatic Adjustment. The Early Pliocene intertidal 171 units surveyed at Camarones were subject to processes that 172 caused their past and current elevation to depart from eustasy. 173 These processes must be accounted for in order to reconstruct 174 global mean sea level at the time of formation. We calculate 175 Glacial Isostatic Adjustment (GIA) using 36 different Earth mod-176 els. For this site, we calculate a GIA correction of −14.6 ± 3.2m 177 (1σ) (see Methods for details). This value is subtracted from the 178 Preprint -An Early Pliocene relative sea level record from Patagonia (Argentina) 4 observed paleo RSL and the uncertainty propagated. This correction is a combination of effects associated with the ongoing 180 response to the last deglaciation and Antarctic ice sheet oscil-181 lations during the early Pliocene 2 . The former contribution is 182 −9.5 ± 3m (1σ), which means that the Argentinian coast today 183 experiences sea level fall due to a combination of effects associ-184 ated with postglacial rebound due to the melting of the glacial 185 Patagonian ice sheet as well as continental levering, ocean sy-186 phoning, and rotational effects. Once fully relaxed, sea level 187 at Camarones will therefore be lower (and a paleo sea level 188 indicator higher) by approximately 9.5m than it is today. The ad- these DT models average to 4.5 ± 2.2m/Ma (Table 3). Account-208 ing for the age of the deposit, this leads to a downward correction 209 of our global mean sea level inference by 22.5 ± 11.0m (1σ).

210
As is apparent from the variation of estimates for the dynamic 211 topography rate, this correction remains quite uncertain and the 212 true value can possibly be even outside of this range given that 213 it is difficult to fully explore model uncertainties. While uncertainties in the estimated vertical land motions nec-237 essarily lead to large uncertainties in the global mean sea level 238 estimates, there is overlap between the calculated global mean 239 sea levels for Camarones (28.4 ± 11.7m, 1σ) and Coves d'Artá 240 (25.1m, with 16 th -84 th percentiles of 10.6-28.3m, Figure 3B).

241
An estimate of global mean sea level from the proxy record at 242 Cliffs Point, South Africa 10 is characterized by greater uncer-243 tainty. Corrected with the same GIA models used for Camarones 244 (Table 2), this data point indicates a paleo RSL at 44.7 ± 2.7m 245 (1σ). The same DT models used at Camarones indicate possible 246 uplift of 3.4 ± 6.3m/Ma (1σ). This results in an average global 247 mean sea level estimate that aligns with that from Camarones, 248 but bounded by very large uncertainties ( Figure 3B).

249
Despite the relevant uncertainties, the average global mean sea 250 level calculated from the geological facies reported in Argentina 251 (this study), South Africa 10 and Spain 9 is well above modern 252 sea level. In each area, post-depositional uplift contributes sig-253 nificant uncertainties to these estimates. We remark that, within 254 each of these broader regions, there are other well-constrained 255 Plio-Pleistocene sea level index points that may eventually pro-256 vide a better calibration for modeled uplift rates.   Note that this setup to calculate the GIA correction deviates 400 slightly from the one described in Dumitru et al. 9 in three small 401 ways, (1) we only consider one GMSL history for the Pliocene 402 rather than a range of histories, (2) we only consider variability 403 in southern hemisphere ice sheets and (3) we calculated GMSL 404 as described above rather than as changes in grounded ice vol-405 ume.

406
The GIA corrections from both processes are combined. In a 407 last step we consider the age range for each sea level indicator 408 and average the GIA correction during warm periods, which we 409 define as times that had higher than average sea level over this 410 time period 9 . The mean and standard deviation that is obtained 411 is shown in table 2. We also show the GIA correction calculated 412 in 9 and note that the difference in mean GIA estimates stems 413 mostly from our different definition of global mean sea level. 414 For the analysis in the main text we use the GIA correction 415 described in 9 for the datapoint on Mallorca and not the one 416 recalculated here.    Figure 1 and Figure 3 were downloaded from 452 GitHub (https://github.com/fraxen/tectonicplates/blob/master/). 453 We

46
Marine Isotopic Stage 5e. The Last Interglacial is also pre-47 served in the form of relic beach ridges in the Camarones 48 area. These were studied by different authors throughout the 49 years 9;12;10;23 , and were dated to MIS 5e using Electron Spin 50 Resonance and U-Series on mollusks (Supplementary Table 1). 51 A recent study by Pappalardo et al. 2015 9 provides more pre-52 cise measurements, interpretations and additional dating of the 53 MIS 5e beach ridge complex at Camarones. According to these 54 authors, the MIS 5e beach ridges at Camarones were formed in 55 correspondance with a paleo RSL at 7.5 +2/-3.5m above present. 56 Marine Isotopic Stage 11. At one site south of Camarones 57 town, articulated shells from (Sample Pa 35) was dated by Schell-58 mann and Radtke (2000) 12 as MIS 9 or older. U-series mollusk 59 ages by Pappalardo et al. (2015) 9 confirm the attribution to 60 MIS 11. We measured the deposits dated by these authors at 61 16.7 ± 0.4m above present sea level.

62
Detailed description of Roadcut and Caprock units 63 at Camarones

64
The Roadcut section (Supplementary Figure 1) is characterized 65 by the bedrock (Río Chico formation) outcropping from the road 66 level up to ca.12m above it, mostly sheltered by a tick debris. 67 The topmost part of the bedrock is exposed for a maximum 68 thickness of 1.2m in the western part of the outcrop and it is 69 shaped as a flat, gently eastward (i.e. seaward) dipping platform. 70 All the overlying units are separated from it by a sharp erosional 71 unconformity. Less than 1 km south of the Roadcut, another 72 outcrop shows the same geological context. We refer to this as 73 the Caprock outcrop (Supplementary Figure 2). This rests on 74 a relative topographic high of the bedrock, which in this point 75 is represented by the volcanic Complejo Marifil, capped by a 76 thin sedimentary unit, as thick as 1m maximum, identical to the 77 upper part of the Cp Unit observed in the Roadcut section. Each 78 overlying unit is described separately hereafter.   Table 3).   Table 3) and a scoring system from 216 "1" to "3" was established based on long-term uncertainties of